Arbitration Innumeracy

Abstract

Arbitration innumeracy, as I use the phrase here, is the “inability to deal comfortably with the fundamental notions of number and chance” in evaluating arbitration, particularly consumer and employment arbitration. This article discusses a number of examples of possible arbitration innumeracy — cases in which statistics about arbitration are incomplete or outdated, misunderstood or misused. In particular, it examines empirical studies on: The use of arbitration clauses in credit card agreements; Outcomes in consumer arbitration; Arbitrator selection by the National Arbitration Forum; Class arbitration waivers in consumer arbitration clauses; The incentives of arbitrators and repeat-player bias; and Unintended consequences of restrictions on consumer and employment arbitration clauses.

Each of these topics illustrates a different issue of arbitration innumeracy, ranging from samples that are not representative of the population as a whole to comparisons that do not compare like cases. The sections of this Article put the issue in more complete empirical context and discuss briefly how the arbitration innumeracy impacts the policy debate over consumer and employment arbitration.

For consumers of empirical data, avoiding arbitration innumeracy requires a sensitivity to basic statistical concepts and a willingness to look skeptically at empirical research, even when it confirms one’s previously held views. For producers of empirical data, avoiding arbitration innumeracy requires a willingness to apply proper empirical techniques and, importantly, to recognize the limitations of one’s data. As empirical data becomes ever more important to the intensifying debate over consumer and employment arbitration, avoiding arbitration innumeracy is essential to making sound public policy.